Analysis
Jensen Huang used a CNBC Mad Money appearance to directly rebut a specific and growing line of criticism: that Nvidia's expanding financial support for AI labs -- investments, credit facilities, and compute financing arrangements -- amounts to vendor financing that inflates the company's own reported growth, CNBC reported. Huang's core argument: the risk to Nvidia is low because its GPUs are a redeployable asset -- if a company Nvidia financed fails or scales back, the compute infrastructure underlying that deal can be reallocated to other paying customers, unlike a dot-com-era loan that simply disappears if the borrower goes under.
The comments came the same day Nvidia reported blockbuster Q2 fiscal 2027 results: revenue of $96.2 billion, up 106% year-over-year, with guidance of $108 billion for the current quarter. Huang went further on the earnings call, forecasting 70% revenue growth for fiscal 2028 -- a number he said reflects supply constraints rather than the ceiling on actual demand, telling investors 'even though our demand is much greater than 70%, our supply allows us to confidently deliver 70%.'
Room for disagreement: the redeployability argument assumes Nvidia's GPU supply stays scarce enough that any capacity freed up by a failing customer gets immediately absorbed by the next one in line -- true today, given demand is outrunning supply by Huang's own account, but a much weaker defense in a scenario where AI infrastructure demand cools broadly rather than just at one struggling lab. If multiple financed customers stumbled at once in a softer demand environment, the 'redeploy the hardware' safety valve gets a lot less reassuring, because there'd be fewer buyers competing for that freed-up capacity.
It's also worth separating two different things Huang is defending in one answer: financing that helps a customer buy Nvidia chips it would otherwise have bought with its own capital (genuinely low risk, mostly a working-capital convenience) versus financing that effectively creates the demand being reported as organic revenue growth (the actual circular-financing concern critics are raising, and one Huang's public comments haven't fully separated from the first case).
For VCs and founders in AI infrastructure, the practical takeaway isn't whether Huang is right or wrong in the abstract -- it's that Nvidia's own growth numbers are now partly a function of Nvidia's own capital allocation decisions, which makes reported AI-infrastructure demand growth harder to read as a pure market signal than it was two years ago, when GPU sales more cleanly reflected independent buyer demand.
